Automatic Exploratory Data Analysis Report

AutoML Leaderboard

Best model	name	model_type	metric_type	metric_value	train_time
	1_Baseline	Baseline	logloss	0.693147	2.76
	2_DecisionTree	Decision Tree	logloss	0.296321	37.17
the best	3_Default_Xgboost	Xgboost	logloss	0.0498985	130.3
	4_Default_NeuralNetwork	Neural Network	logloss	0.446357	455.11
	5_Default_RandomForest	Random Forest	logloss	0.236005	140.21
	Ensemble	Ensemble	logloss	0.0498985	19.46

AutoML Performance

AutoML Performance Boxplot

Features Importance

features importance across models

Spearman Correlation of Models

models spearman correlation

Summary of 1_Baseline

<< Go back

Baseline Classifier (Baseline)

n_jobs: -1
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

1.4 seconds

Metric details

	score	threshold
logloss	0.693147	nan
auc	0.5	nan
f1	0.666667	0.45
accuracy	0.5	0.45
precision	0.5	0.45
recall	1	0.45
mcc	0	0.45

Confusion matrix (at threshold=0.45)

	Predicted as 0	Predicted as 1
Labeled as 0	0	35212
Labeled as 1	0	35212

Learning curves

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 2_DecisionTree

<< Go back

Decision Tree

n_jobs: -1
criterion: gini
max_depth: 3
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

35.4 seconds

Metric details

	score	threshold
logloss	0.296321	nan
auc	0.930521	nan
f1	0.667539	0
accuracy	0.50196	0
precision	0.500982	0
recall	1	0
mcc	0	0

Confusion matrix (at threshold=0.0)

	Predicted as 0	Predicted as 1
Labeled as 0	138	35074
Labeled as 1	0	35212

Learning curves

Decision Tree

Tree #1

Rules

if (SizeOfStackReserve > -0.002) and (SizeOfInitializedData > -0.02) and (ResourcesMeanSize > -0.005) then class: 0 (proba: 91.53%) | based on 70,231 samples

if (SizeOfStackReserve <= -0.002) and (ResourcesMeanSize > -0.005) and (CheckSum > -0.21) then class: 1 (proba: 98.87%) | based on 66,811 samples

if (SizeOfStackReserve > -0.002) and (SizeOfInitializedData > -0.02) and (ResourcesMeanSize <= -0.005) then class: 0 (proba: 68.51%) | based on 42,565 samples

if (SizeOfStackReserve > -0.002) and (SizeOfInitializedData <= -0.02) and (CheckSum > -0.21) then class: 1 (proba: 86.79%) | based on 17,393 samples

if (SizeOfStackReserve > -0.002) and (SizeOfInitializedData <= -0.02) and (CheckSum <= -0.21) then class: 0 (proba: 91.87%) | based on 8,283 samples

if (SizeOfStackReserve <= -0.002) and (ResourcesMeanSize > -0.005) and (CheckSum <= -0.21) then class: 1 (proba: 82.14%) | based on 5,224 samples

if (SizeOfStackReserve <= -0.002) and (ResourcesMeanSize <= -0.005) and (SizeOfStackReserve <= -0.002) then class: 0 (proba: 100.0%) | based on 412 samples

if (SizeOfStackReserve <= -0.002) and (ResourcesMeanSize <= -0.005) and (SizeOfStackReserve > -0.002) then class: 0 (proba: 52.12%) | based on 353 samples

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

SHAP Importance

SHAP Dependence plots

Dependence (Fold 1)

SHAP Dependence from Fold 1

SHAP Decision plots

Top-10 Worst decisions for class 0 (Fold 1)

SHAP worst decisions class 0 from Fold 1

Top-10 Best decisions for class 0 (Fold 1)

SHAP best decisions class 0 from Fold 1

Top-10 Worst decisions for class 1 (Fold 1)

SHAP worst decisions class 1 from Fold 1

Top-10 Best decisions for class 1 (Fold 1)

SHAP best decisions class 1 from Fold 1

<< Go back

Summary of 3_Default_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

n_jobs: -1
objective: binary:logistic
eta: 0.075
max_depth: 6
min_child_weight: 1
subsample: 1.0
colsample_bytree: 1.0
eval_metric: logloss
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

127.3 seconds

Metric details

	score	threshold
logloss	0.0498985	nan
auc	0.998071	nan
f1	0.983644	0.494486
accuracy	0.983599	0.494486
precision	1	0.999999
recall	1	1.74021e-09
mcc	0.967213	0.494486

Confusion matrix (at threshold=0.494486)

	Predicted as 0	Predicted as 1
Labeled as 0	34538	674
Labeled as 1	481	34731

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

SHAP Importance

SHAP Dependence plots

Dependence (Fold 1)

SHAP Dependence from Fold 1

SHAP Decision plots

Top-10 Worst decisions for class 0 (Fold 1)

SHAP worst decisions class 0 from Fold 1

Top-10 Best decisions for class 0 (Fold 1)

SHAP best decisions class 0 from Fold 1

Top-10 Worst decisions for class 1 (Fold 1)

SHAP worst decisions class 1 from Fold 1

Top-10 Best decisions for class 1 (Fold 1)

SHAP best decisions class 1 from Fold 1

<< Go back

Summary of 4_Default_NeuralNetwork

<< Go back

Neural Network

n_jobs: -1
dense_1_size: 32
dense_2_size: 16
learning_rate: 0.05
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

452.2 seconds

Metric details

	score	threshold
logloss	0.446357	nan
auc	0.866918	nan
f1	0.807344	0.352627
accuracy	0.78922	0.499792
precision	0.949436	0.822672
recall	1	0.00217924
mcc	0.585203	0.499792

Confusion matrix (at threshold=0.499792)

	Predicted as 0	Predicted as 1
Labeled as 0	25121	10091
Labeled as 1	4753	30459

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 5_Default_RandomForest

<< Go back

Random Forest

n_jobs: -1
criterion: gini
max_features: 0.9
min_samples_split: 30
max_depth: 4
eval_metric_name: logloss
explain_level: 2

Validation

validation_type: split
train_ratio: 0.75
shuffle: True
stratify: True

Optimized metric

logloss

Training time

137.3 seconds

Metric details

	score	threshold
logloss	0.236005	nan
auc	0.967584	nan
f1	0.904597	0.444436
accuracy	0.907319	0.444436
precision	1	0.991008
recall	1	0.00358889
mcc	0.818119	0.478128

Confusion matrix (at threshold=0.444436)

	Predicted as 0	Predicted as 1
Labeled as 0	32953	2259
Labeled as 1	4268	30944

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

SHAP Importance

SHAP Dependence plots

Dependence (Fold 1)

SHAP Dependence from Fold 1

SHAP Decision plots

Top-10 Worst decisions for class 0 (Fold 1)

SHAP worst decisions class 0 from Fold 1

Top-10 Best decisions for class 0 (Fold 1)

SHAP best decisions class 0 from Fold 1

Top-10 Worst decisions for class 1 (Fold 1)

SHAP worst decisions class 1 from Fold 1

Top-10 Best decisions for class 1 (Fold 1)

SHAP best decisions class 1 from Fold 1

<< Go back

Summary of Ensemble

<< Go back

Ensemble structure

Model	Weight
3_Default_Xgboost	1

Metric details

	score	threshold
logloss	0.0498985	nan
auc	0.998071	nan
f1	0.983644	0.494486
accuracy	0.983599	0.494486
precision	1	0.999999
recall	1	1.74021e-09
mcc	0.967213	0.494486

Confusion matrix (at threshold=0.494486)

	Predicted as 0	Predicted as 1
Labeled as 0	34538	674
Labeled as 1	481	34731